Acoustic vs. Electronic Singer-Songwriters
Introduction
Introduction
This assignment will be about the differences and similarities between electronic and acoustic singer-songwriters. Acoustic and electronic music are a world apart, but it seems now that they can be closer than expected by the fact that both acoustic and electronic singer-songwriters exist. I define acoustic singer-songwriters as 1 person that sings and plays exclusively on 1 acoustic instrument, mainly guitar or piano. Examples are Passenger and Ed Sheeran.
Electronic singer-songwriters are defined as people who produce their own music containing (not necessarily but also not limited to) synthesizers, looping stations or drums computers, and are able to perform it live. The music must contain roughly the same amount of singing as the acoustic songs. Examples are Chet Faker and JAIN.
The question that will be examined in this report is: What are the major differences between music of acoustic and electronic singer-songwriters and where do they overlap?
In this investigation, the following analyses will be presented. When 2 songs are compared sections (2 to 5), the first song always is a song from the acoustic playlisy and the second song is one from the electronic playlist. The songs all always archetypes of the playlits. The reason for this is to be able to generalize the conclusions (as much as possible) over the whole playlists.
- Mode, danceability, energy
- Chromagram ‘Dynabeat’ and ‘Will’s lullaby’
- SSM Timbre & Pitches ‘North Swallow’ and ‘Clean’
- Chordograms ‘This’ and ‘LEKKER MET DE MEIDEN’
- Keygrams ‘Heart’s On Fire’ and ‘Doesn’t Matter (voleur de soleil)’
Corpora
There are two corpora that will be compared with each other. These are Spotify playlists. For the acoustic singer-songwriters, a playlist generated by Spotify that was called: “Acoustic Singer-songwriters” was chosen. The playlist was checked manually and all the songs satisfied the criteria stated above. This corpus represents the acoustic singer songwriters pretty well. There are 233 songs in this playlist.
The playlist for the electronic singer-songwriters was made by a Spotify user and complemented by me. It contains many songs that fulfill the criteria from above, but also some songs that belong to the indie genre, with lots of electric guitars, but no electronic instruments. There are 227 songs in this playlist. These songs were manually filtered from this playlist as much as possible.
Mode, energy, danceability
The biggest difference between the two playlists is the mode of the songs. In the acoustic playlist, 6.87 percent of the songs is in minor key. In the electronic playlist, 32.6 percent of the songs is in minor. That is a large difference. That difference is displayed in the barplots shown in this graph.
The difference in danceability is also visible in the graph. There is not one song on the acoustic playlisy that has a danceability larger than 0.8. In the electronic playlist there are many songs that exceed that threshold.
The mean danceability of the acoustic playlist is 0.55, (SD = 0.11). and the mean danceability of the electronic playlist is 0.62, (SD = 0.16).
More electronic songs are speechy than aocustic songs. Only 2 acoustic songs have a speechiness higher than 0.1, but the electronic songs have a lot more speechy songs.
The mean speechiness of the acoustic playlist is 0.04, (SD = 0.01). and the mean speechiness of the electronic playlist is 0.07, (SD = 0.07).
The mean energy of the acoustic playlist is 0.38, (SD = 0.16). and the mean energy of the electronic playlist is 0.51, (SD = 0.18).
The higher energy of the electronic playlist could be explained if the tempo of the electronic songs was higher than the tempo of the acoustic songs, but that is not the case, surprisingly. The mean tempo of the acoustic playlist is 116.73, (SD = 30.22). and the mean tempo of the electronic playlist is 119.89, (SD = 27.65).
From this section, we can conclude that electronic songs tend to be more moody than acoustic songs, due to the number of minor songs, acoustic songs are less danceable than electronic ones and contain less speech.
Chromograms ‘Everything’ll Be Alright’ and ‘Dynabeat’
The chromagram of ‘Everything’ll be Alright’ is overall darker than the chromagram of ‘Dynabeat’. The blue rectangles are wider than the other chromagram. That probably means those notes are hold longer than when the rectangles woudl be less wide. Longer notes could cause a lower energy. The energy of this song is 0.157 and the tempo is 170. The tempo measure seemed very high, so the tempo of this song was checked manually by tapping along on the website: http://www.beatsperminuteonline.com/. The outcome was 85 bpm. This is exactly half of the tempo Spotify measured and it makes sense, since one could tap twice as fast and then get 170 bpm as the tempo. Since most other acoustic songs has a tempo around 90 bpm, it seems that 85 bpm would be the correct tempo for ‘Everything’ll be Allright’.
The ‘Dynabeat’ chromagram has more blue than the previous one. That could be because more instruments are used at the same time causing more notes to be played at the same time. You could discriminate different sections in the song. It is also clear that the rectangles are very narrow, which could increase the energy and or tempo. The energy of this song is 0.613 and the tempo is 110.
Chroma- and Timbre Self Similarity Matrices (‘North Swallow’)
The chromabased SSM of ‘North Swallow’ has a very clean checkerboard structure. The intro is used during the verses, which you can clearly see in the first checkerboard block. The bright yellow stripe is the first chorus, which is repeated 2 more times. The verses and chorusses are repeated with exactly the same sounds, which could be reason this SSM is so clean and tidy.
The timbre based SSM is not very bright because the song contains mainly acoustic guitar. There are barely any changes in the instruments that are used. There are drums, but they are roughly the same during the song. The timbre SSM of ‘Clean’ is also less bright than the chroma SSM, but it is a bit brighter than this timbre SSM.
Chroma- and Timbre Self Similarity Matrices (‘Clean’)
The chroma- and timbrebased Self Similarity Matrices (SSM’s) line up really well. The song has a chorus of 2 sentences and 2 verses. The chorusses are shown clearly in the three yellow lines at the beginning of the song. Later on, the lines appear again but they are surrounded by other yellow parts. This proabably is the part where 2 voices sing different parts at the same time, which causes the chorus to stand out less in this SSM. The chromabased SSM is overall very bright and yellow. It is brighter than the chromabased SSM of ‘North Swallow’. We’ve already seen this difference between the chromagrams of ‘Everything’ll be Alright’ and ‘Dynabeat’. The differnce between the SSM’s could as well explained by the fact that electronic musicians can use more instruments at the same time by looping them.
The chorusses are visible in the timbre SSM (the yellow vertical stripes), but the timbre of the chorusses are different. The instrumental music during the chorus changes almost everything it is sung, that is why the timbre is not exactly the same, but since it is the same part, you can still distinguish the chorusses in this SSM.
We can conclude that the acoustic song, ‘North Swallow’, has a clearer chroma structure than the electronic song, ‘Clean’, but that ‘Clean’ has more timbre variation.
Chordograms ‘This’ and ‘LEKKER MET DE MEIDEN’
‘This’ has roughly the same chords for the whole song, mostly B:maj/B:7 and E:maj/E:7. You can see this in the chordogram by the blue parts at the top. The song is pretty monotonous and doesn’t change much.
‘LEKKER MET DE MEIDEN’ uses more different chords, but still the same chords over de whole song. Different sections are only distinguished by the yellow parts, but the chords that are played, are played the whole song. The chorus sound different than the verses, but it could be that the key of the chorus is the same as the verses and the same chords are just played differently.
To investigate this further, I have also made keygrams of these to songs, show in the net tab.
Keygrams ‘This’ and ‘LEKKER MET DE MEIDEN’
In the keygram of ‘This’, one can see the different sections more clearly. The keygram is ofcourse aligned by sections, so the differences are clearer already, but you can now see two sections in the keygram around 100 and 150 ms that are different than the rest. These sections are indeed different than the rest when one listens to the song. At closer inspection of the chordogram, one can see these differences, but they’re a lost less clear than in the keygram.
The keygram and chordogram of ‘LEKKER MET DE MEIDEN’ is much more the same. The sections in the keygram are very visible in the chordogram as well. The key changes in the second section, which was not what was expected based on the chordogram.
Keygrams ‘Heart’s on Fire’ and ‘Doesn’t Matter (voleur de soleil’
These keygrams aren’t all that different. ‘Doesn’t Matter’ has a large section where a lot of keys match the pitch information, whereas ‘Heart’s On Fire’ most of the time has one or two clear dark blue lines. This difference could again be explained by the use of multiple instruments at the same time in the electronic songs.
Large Graph NO?
Since we’re going to expand these kind of graphs next week, I will write down commentary and conclusions for this graph, that might next week.
Timbre Coefficients Of the Playlists
The biggest timbre differences seem to be in coefficients c3, c4 and c5,
Classification results with confusion matrix
With the decision tree made a prediction with an accuracay of 0.8586498. The most distinctive features, according to the model, are danceability, and timbre features c08 and c06.
Speechiness is very high as well. In the graph I made earlier in the project, I’ve changed the energy parameter to speechiness.
Randomforests
Again, the most distinctive features are danceability, and timbre features c08 and c06.
Week 12.1 Acoustic
Week 12.1 Electronic
12.2 Acoustic
12.2 Electronic
12.3 Acoustic
12.3 Electronic
12.4 Acoustic
12.4 Electronic
12.5 Acoustic
12.5 Electronic
12.6 Acoustic
K-means clustering with 4 clusters of sizes 82, 75, 59, 17
Cluster means:
danceability energy loudness speechiness acousticness
1 0.5945625 0.2259876 0.4528687 0.12363778 0.8572239
2 0.6423976 0.4349278 0.7235099 0.09130024 0.7585877
3 0.5958016 0.6494009 0.7752260 0.11630004 0.2554370
4 0.5098759 0.3743457 0.6280430 0.05824990 0.7037131
instrumentalness liveness valence tempo duration_ms C
1 0.084058161 0.1173766 0.3095523 0.4305745 0.3875706 0.5809963
2 0.008981131 0.1623495 0.4444734 0.4062712 0.3150036 0.5911567
3 0.033049142 0.1440705 0.4634156 0.4923842 0.4098423 0.5460951
4 0.040494389 0.1186221 0.3029339 0.4742092 0.4130989 0.5166666
C#|Db D D#|Eb E F F#|Gb G
1 0.4608865 0.4372440 0.2398491 0.6266126 0.4315110 0.4532916 0.5490975
2 0.4708887 0.4588901 0.2608821 0.6390074 0.3721792 0.4190061 0.5188606
3 0.5501105 0.4039312 0.2885551 0.5981444 0.4165717 0.4507213 0.4348392
4 0.3440321 0.3106392 0.7219037 0.2520947 0.3898044 0.3546851 0.3330570
G#|Ab A A#|Bb B c01 c02 c03
1 0.3342182 0.6592986 0.2800221 0.4587963 0.3639333 0.4099251 0.3807794
2 0.3376687 0.7276764 0.2953808 0.4288439 0.6590206 0.5969930 0.5405418
3 0.3793884 0.6897760 0.3355625 0.4086013 0.7451312 0.7488870 0.6169783
4 0.6510080 0.4924303 0.6467099 0.3064447 0.5761834 0.5105540 0.4902663
c04 c05 c06 c07 c08 c09 c10
1 0.3821068 0.4536401 0.3884925 0.3933494 0.5378442 0.6152949 0.5820665
2 0.5069806 0.5494442 0.4362977 0.4081892 0.5976034 0.5193171 0.5352854
3 0.6104546 0.4377414 0.2906470 0.5972144 0.6640087 0.5140289 0.6170355
4 0.4444098 0.5002168 0.2981934 0.4646175 0.6293079 0.5922381 0.5735089
c11 c12
1 0.5501535 0.4646091
2 0.5707052 0.4326138
3 0.5864885 0.3571298
4 0.6061701 0.4254055
Clustering vector:
Heart's On Fire Only Love The Flood
1 3 2
Should Have Known... Hot Scary Summer Olivia
1 2 2
Never Come Back A... Ghosts Kathleen
3 3 3
Beautiful Girl Masterpiece - (solo) Fletcher Bay
1 2 2
Boy Turns Around the ... You Remind Me
2 3 3
Come Alive Wash Away (reprise) Cautionary Tale
1 2 3
My Heart Is With You Dirty Old Town Changes
1 3 2
Fall Like a Feather Happiness Can You Hear It
3 3 3
Heart and Soul Brighter Shade of... Hard It Goes
3 2 3
Your Love Fellow Man Hollering Hearts
1 1 3
Dark Child - Sing... No One Says You H... Secrets (Cellar D...
4 2 3
Frozen River God Damn Fool Sweet Madeline
2 2 3
Inheritance Steady My Bones California
3 3 2
Last Night Just Like Anythin... Moonshine
1 1 1
Misfits My Repair Damn Sure
2 3 1
It's a Job Let the Meter Run Americana Dream
3 2 1
Good Times Knocki... A Part Life Is A Song
2 1 2
Boxes We Were Meant to ... Wasn't Expecting ...
1 2 2
The Dusty Roads We Tried Like Hell Chrissie Hynde
1 2 3
Leaving Charm City Keep On, Keepin' On Stop Your Tears
2 3 4
Fireflies Alright, Alright Whale Song
2 3 1
Harmony, CA Porch Light Way Out
2 2 1
After Rain Head over Heels Home
4 3 2
Bravery Folk Bloodbath It Ain't Over Yet
2 1 3
The Great and Und... The Ballad Of Jim... Little Boy
2 1 3
Take This Heart o... Jump to Mexico Bright Side
1 1 2
Hope vs. Pray Alstroemeria The House of the ...
1 1 4
Shooting Star A Few Miles More North Swallow
3 2 3
By the Ocean Do You Need Me Near Hollywood's Heroes
2 1 1
Half Moon Lane Blues Run The Game Sweetly
2 1 1
Change It All All the Small Things Even the Darkness...
1 1 2
City by the Sea Sister California Grow Up
2 2 3
Friend and a Friend Up from the Ashes Now That I Know
4 4 1
My Jealous Heart Figure It Out How Soon The Dawn
3 3 2
Green and Blue Call It Dreaming Foreign Hands
1 2 3
Fallen so Young Bear Paper Football
2 4 3
Beth C'est la vie - Ac... The Middle
2 2 2
Simple Song Nice And Quiet I Know
2 1 2
Everything Is Mad... Room with a View Disappears With Time
3 1 4
Each Time You Call An Island Might N... All I Need
4 3 3
Destitute I Was Just Leaving Old Town
3 1 1
In My Skin The Push and Pull Common Man
1 1 2
Fire on Bowery (A... Almost Home The One That Make...
1 2 3
I Can't Stand the... Oh Love Opossum
2 3 2
The Weary Kind Driftwood Chicago Song
1 3 3
Sideways Gold Pour out Love
2 2 3
In My Arms Down by the Water Carry On
1 2 2
Do Something Cry for a Sparrow Eye on the Sparrow
3 1 3
Under Control Karma Echoes in the Wind
1 3 1
Hey Louise Where All the Tim... Carry the Blame
2 1 1
Blood Drive Euphoria Flowers
1 3 4
St. Vincent De Pa... Gauntlet in the Sun I Got You, Honey
3 3 1
Georgia Hey, Hey, Hey Bow And Arrow
1 2 2
The Back Seat Try to Believe Mountain Girl
1 2 1
Hey Old Friend Emily Golden (Acoustic)
1 1 1
Daddy's Little Girl Don't Love Anybody Illinois
1 2 4
Born into a Ballroom Friends Again The Cold
2 1 1
More Than I Can C... Climb Find Your Tribe
1 3 3
The Ballad of Don... Chemicals Rearranged
2 1 1
The Cold Side I Knew I'd Find You Don't Think Twice...
3 1 2
All Your Love with you/for you City Folks
1 1 3
Constellations Laughing River Headlights on the...
2 1 2
Hold My Name Darlin It’s Alright
1 3 2
Daisy Ten Little Indian... I'll Be Around
2 3 1
Cityscapes Give Me Dreams Clean Eyes - Acou...
2 1 4
Beautiful The Edges - Acoustic Best I'll Ever Sing
4 2 2
Change (feat. Jam... Call on Me Everything'll Be ...
2 2 1
This Pink Moon Pure
1 1 1
This Is Why I Nee... Sophie Altruistic Holiday
1 4 1
New Morning Strangers Old Box
1 1 2
Ghost Writer The Drying of the... Blue Sky Fall
2 2 3
River Mouth Permanent Way Cold Fronts (Acou...
3 1 1
The Pool Untitled (Love Song) Where You Stay
1 1 1
Firewood The Wolves Everything Falls ...
4 1 2
Siren Song Hey, Runner! Maggie Cassidy
4 2 1
Our Remains Ran Away Black Hair
4 3 2
Eraser Holy Road Gotta Live It
3 1 2
Halo Reverie
2 3
Within cluster sum of squares by cluster:
[1] 72.17928 53.95381 44.49038 13.25162
(between_SS / total_SS = 24.1 %)
Available components:
[1] "cluster" "centers" "totss" "withinss"
[5] "tot.withinss" "betweenss" "size" "iter"
[9] "ifault" 12.6 Electronic
K-means clustering with 4 clusters of sizes 57, 80, 22, 68
Cluster means:
danceability energy loudness speechiness acousticness
1 0.7508322 0.5738552 0.7324896 0.19722640 0.2915702
2 0.4198558 0.3691684 0.5484252 0.05753439 0.7059050
3 0.4648019 0.5690536 0.5978975 0.03139009 0.5233901
4 0.5234351 0.7398356 0.8099765 0.06184563 0.1773399
instrumentalness liveness valence tempo duration_ms C
1 0.05038170 0.1423523 0.5992264 0.3299998 0.1899365 0.3472847
2 0.04297441 0.1391188 0.2231180 0.3972877 0.2340116 0.4372124
3 0.78768700 0.1795437 0.3039077 0.3461266 0.3249897 0.4062075
4 0.06200915 0.2082672 0.4683283 0.3842839 0.2509005 0.3827292
C#|Db D D#|Eb E F F#|Gb G
1 0.4092647 0.4175862 0.5767464 0.4617427 0.4982964 0.4080781 0.4145160
2 0.3803276 0.4929994 0.5171023 0.5358861 0.5630530 0.3622280 0.4617332
3 0.3700269 0.4466878 0.4850679 0.5809377 0.4862175 0.3436919 0.5716679
4 0.4497869 0.4840818 0.5153932 0.4846235 0.5081597 0.3894674 0.4256763
G#|Ab A A#|Bb B c01 c02 c03
1 0.4787478 0.4722760 0.4703564 0.4516678 0.6415198 0.6356913 0.5026028
2 0.3519385 0.4800587 0.3990816 0.4163068 0.5091024 0.4035093 0.4937575
3 0.4344616 0.5540102 0.3345721 0.4170378 0.5943481 0.3481488 0.5831986
4 0.4377335 0.5341719 0.4090992 0.4033880 0.7843318 0.6755653 0.5665470
c04 c05 c06 c07 c08 c09 c10
1 0.6863224 0.3954816 0.5226883 0.4919488 0.5471060 0.4540640 0.3953432
2 0.3376895 0.4403062 0.3384194 0.4115694 0.5076199 0.5080892 0.2881285
3 0.5451991 0.3494082 0.2027895 0.6008378 0.4713734 0.6240840 0.3245671
4 0.5039098 0.3492484 0.2947211 0.5184929 0.4810659 0.4734508 0.3372052
c11 c12
1 0.5050898 0.4509733
2 0.6996649 0.4377992
3 0.7811249 0.4672941
4 0.6924676 0.4218385
Clustering vector:
Today Why Don't You Listen Conrad
4 4 4
Rennen The Wheel Signal
2 1 1
Wildfire Hold On 22 (OVER S∞∞N)
1 3 2
33 “GOD” 8 (circle) Get Low
2 2 2
Drop The Game Gold Talk Is Cheap
4 2 4
1998 Cigarettes & Lone... Someone That Love...
1 4 4
Good Together Hurt Me Falling Short
4 4 2
Love Is Blind Cliff Without
4 2 1
Retrograde Choose Me On Hold
2 3 4
Taro Ms Fitzpleasure
2 2 4
Sharpness Night Air Shoulda
1 1 2
Cirrus Steadier Real Love Baby
3 2 4
Dreams 17 Kingdom Leader
2 3 4
Cannons Afternoon Oh Jordan
3 3 2
Landscape Escape Young Knife Jailbird
2 2 2
The Sailor's Bonnet The Fiddlers Nothing, Nobody, ...
2 2 1
Fool The Spinning Wheel Chewing Gum (Conc...
1 2 2
Song For You My Bed Looking Too Closely
1 1 3
Wonderful Life Welcome Home, Son Old Friend
2 3 4
Hang a String of ... Swimming Away Beach House
4 4 2
Gooey Youth Life Itself
2 4 4
Coffee Radio Hey Mami
2 4 1
Digital Witness Cruel Cheerleader
1 4 2
Strange Mercy Two Weeks Yet Again
4 4 4
Here with Me Nitesky feat. Joh... Care
2 3 3
Human Jacques Cousteau Airport
1 2 4
Dog Days Are Over You've Got The Love I Went Too Far - ...
4 4 4
Running With The ... I Will Wicked Games - Ra...
4 4 4
Sorry Maar Wie Van Ons Is Blind Ik Loop
1 1 2
Liefde Van De Vloer Club Overleeft Wel Murder to the Mind
1 2 4
To Death With You Bad Behaviour Tell me what you'...
2 1 4
Catherine the Wai... 29 #Strafford APTS Steal
2 2 2
Midas In The Waiting Line Sunday Interlude
1 2 4
Quite Like Rose Rouge Heart Shaped Box
1 4 2
The Invincible Sidesteppin' Love in Music
1 1 4
From the Tree Phoenix - Little ... Born To Be Yours
1 1 4
Marilyn Monroe Hubris Soul Syncable
2 2 4
Come Heads Up Mr Johnson
1 1 1
Lil Mama Hope All my Days
1 1 2
Hob Makeba You Can Blame Me
1 1 1
So Peaceful On My Way Flash (Pointe-Noire)
4 1 4
Alright Oh Man Inspecta
1 1 1
Dream Star Feel It
1 1 4
Abu Dhabi Souldier I Need A Forest Fire
4 1 3
Radio Silence Open Feel Your Weight ...
2 2 3
Waste The Fall Last Dance
2 3 1
Scheef Mee Wacht
2 1 2
Luister Staan Jong
2 1 2
Wakker Stof Naartoe
2 3 2
kom terug Bagagedrager Ik Wil Alleen Maa...
4 4 4
Trein Vuur Dageraad Oostende Berlin
3 4 4
Girlfriend Bad Blood Maybe You're the ...
2 2 4
Follow My Girl Lilo Landslide - Recor...
4 2 2
Saw You in a Dream Somebody you found Swim Against the ...
4 2 2
Good side in Clean Cool Blue
2 2 2
Letter by the Water Sugar Pill 3/3
2 2 2
Count to Nine Face Like Thunder Leon
2 4 2
Pools to Bathe In Teeth Still
2 3 2
Sister We Talk all the Time Comme si on s'aimait
2 4 4
La marcheuse Doesn't matter (v... Goya ! Soda !
1 1 4
Follarse Machin-chose Bruce est dans le...
4 2 1
Le G Les yeux mouillés L'étranger (voleu...
1 2 1
Girlfriend (feat.... 5 dollars Gap in the Clouds
1 1 2
LEKKER MET DE MEIDEN BORDERLINE Breezeblocks
1 1 4
Every Other Freckle Dynabeat Darkest Hour
4 1 3
The Inside Amandine Insensible That Other Girl
2 2 2
Sirens of the Cas... Backseat Love Nikes
1 1 2
Pink + White Solo Nights
4 2 1
White Ferrari Pyramids I Like That
2 4 4
Make Me Feel Pynk (feat. Grimes) Uncatena
1 1 2
The Silicone Veil Able Mable Silver
4 4 4
Glowing Mouth Young Sensualists TURN AWAY
4 4 4
Prince Johnny Winter Blind I Believe In Action
4 2 3
Palace One Thing The Other Side
2 3 3
Apply Also Frightened I Never Learnt To...
3 4 2
Peace Of Mind Rock My Boat
2 3
Within cluster sum of squares by cluster:
[1] 47.63762 70.36262 21.14729 50.16059
(between_SS / total_SS = 24.7 %)
Available components:
[1] "cluster" "centers" "totss" "withinss"
[5] "tot.withinss" "betweenss" "size" "iter"
[9] "ifault" 12.7 Acoustic
[1] 432.0 345.612.7 Electronic
[1] 432.0 345.6Conclusion
Conclusion
- Electronic more minor songs
- Electronic higher danceability and energy
- Electronic songs have more pitch classes in the songs
- Acoustic songs have a clearer structure in chordprogressions and keys.
- Electronic songs have more timbre variation
- The most distinctive features of the playlists are danceability, speechiness and timbre features c06 and c08.
Suggestions for improvement
- Automatically compute percentage of minor songs.
- Write conclusion